HANDLING MULTIPLE FAILURES IN PROCESS NETWORKS

Abstract

Process network is a computation model used in digital signal processing. This model is effective for solving various tasks, which can be divided into multiple concurrent processes. These processes communicate through unbounded FIFO buffers and can be executed simultaneously. In real time digital signal processing applications execution time is often infinite. However, various hardware/software failures can occur. These failures can be single or multiple. In our work, dynamic run-time reconfiguration is introduced into process network which ensures single or multiple error handling, avoiding deadlocks, continuous and on-time result delivery. In this paper, we extend Fault Tolerant Process Network with algorithm for avoiding multiple simultaneous failures, which is also suitable for avoiding a single failure.